Revista de Biología Tropical, ISSN: 2215-2075, Vol. 72: e56487, enero-diciembre 2024 (Publicado Abr. 03, 2024)

Correlation abundance networks for analyzing

biological interactions during cyanobacterial blooms

Florencia Soledad Alvarez Dalinger1, 2; https://orcid.org/0000-0001-9453-3569

Claudia Borja1; https://orcid.org/0009-0004-0900-5374

Liliana Moraña1; https://orcid.org/0000-0001-7445-3537

Verónica Laura Lozano1, 2*; https://orcid.org/0000-0001-8960-9688

1. Facultad de Ciencias Naturales, Universidad Nacional de Salta, Salta, Argentina; floralvarezdalinger0@gmail.com,

claudianborja@gmail.com, lilymorana@gmail.com, vlozano@ege.fcen.uba.ar

2. Centro Científico Tecnológico Salta y Jujuy, CONICET, Argentina; vlozano@ege.fcen.uba.ar (*Correspondance)

Received 27-IX-2023. Corrected 21-II-2023. Accepted 21-III-2024.

ABSTRACT

Introduction: Blooms of cyanobacteria are becoming increasingly common, and understanding their dynam-

ics can be crucial for proposing appropriate management strategies. While physical and chemical parameters

influencing blooms have been widely studied, less attention has been paid to the susceptibility of biological

communities.

Objective: To analyze phytoplankton abundance networks during cyanobacterial blooms at different intensity

levels and how they interact and/or affect the phytoplankton community.

Methods: We used 22 samplings conducted in El Limón reservoir located in Northern Argentina, known for

recurrent cyanobacterial blooms. Each sampling was classified into four levels based on cyanobacteria abundance

(cells/ml): Level 1 (10 000-30 000); Level 2 (30 000-50 000); Level 3 (50 000-100 000); and Level 4 (> 100 000).

For each level, abundance correlation networks were constructed considering all species.

Results: A pattern of decreasing statistically significant abundance correlations was observed as bloom intensity

increased: 219 correlations at Level 1; 144 at Level 2; 80 at Level 3, and only 33 at Level 4. Blooming cyanobacteria

showed few correlations with other species at all levels, indicating a certain independence from the community.

An increase in bloom intensity appears to disconnect the phytoplankton abundance correlation network.

Conclusion: The analysis of abundance correlation networks should be a valuable tool for understanding the

dynamics and development of cyanobacterial blooms, as well as identifying key species in this process.

Key words: reservoir; phytoplankton; Argentina; ecology; management.

RESUMEN

Redes de correlación de abundancia para analizar interacciones biológicas

durante proliferaciones de cianobacterias

Introducción: Las proliferaciones de cianobacterias están volviéndose cada vez más comunes, y comprender su

dinámica puede ser crucial para proponer estrategias de gestión adecuadas. Si bien se han estudiado ampliamente

los parámetros físicos y químicos que influyen en las proliferaciones, se ha prestado menos atención a la suscep-

tibilidad de las comunidades biológicas.

Objetivo: Analizar las redes de abundancia de fitoplancton durante las proliferaciones de cianobacterias a dife-

rentes niveles de intensidad y cómo las mismas interactúan y/o afectan a la comunidad de fitoplancton.

https://doi.org/10.15517/rev.biol.trop..v72i1.56487

AQUATIC ECOLOGY

2Revista de Biología Tropical, ISSN: 2215-2075 Vol. 72: e56487, enero-diciembre 2024 (Publicado Abr. 03, 2024)

INTRODUCTION

Harmful proliferations of algae and cyano-

bacteria are globally recognized for their econo-

mic, health and environmental impacts; and the

overall degradation of resource quality (Kudela

et al., 2017). These are an increasing pheno-

menon, and their occurrence is consistently

expanding into new areas (Cheung et al., 2013).

In the case of cyanobacterial blooms (CB),

which could be associated with toxin release,

their risk assessment continues to be a challen-

ge. The World Health Organization establishes

guideline levels for cyanobacteria based on cell

concentrations (World Health Organization,

2003), which serve as a basis for cyanobacterial

risk assessment. However, most countries lack

specific regulations and pay attention when it

is too late.

The relationship between CB and abiotic

factors is well-documented since it has been

the focal point for researchers over numerous

years. Investigations have primarily concentra-

ted on local abiotic factors such as temperature,

light conditions, nutrient concentrations and

their ratios, pH, and conductivity (O’Neil et al.,

2012; Paerl & Otten, 2013; Paerl & Otten, 2016).

However, the impact of these events on the rest

of the phytoplankton community and biologi-

cal parameters is much less understood, and

little is known about possible early biological

warnings. It has been previously postulated that

the establishment of CB could be associated

with other interactions within the community

beyond nutrient competition (Kokociński et al.,

2021) but this biological standpoint has been

poorly explored until today.

Co-occurrence and/or correlation abun-

dance networks emerge as intriguing tools to

comprehend species interactions (D’Amen et

al., 2018). These networks could unveil interac-

tion patterns and novel insights into potential

connections among various species (Berry &

Widder, 2014). In general, they are employed

to explore how microbial interactions respond

to environmental disturbances and are widely

used in microbiome research (Lu et al., 2022)

but have been little used in phytoplanktonic

communities. Lozano (2022) has explored the

possibility of using correlation abundance net-

works as an early tool for assessing the impact

of herbicidal in freshwater ecosystems with

promising results. Biological interactions could

be manifested in the structure of co-occurrence

networks (Rüger et al., 2021) and, in the case

of correlation of abundance networks, positive

correlation could be associated with taxa acting

similarly, cooperation, facilitation, mutualism,

and symbiosis, while negative correlations

could be associated to competition and/or taxa

presenting opposing behaviors (Lozano, 2022).

Although such networks must reflect these bio-

logical interactions, further research is needed

to verify this correspondence, since correlations

Métodos: Se realizaron 22 muestreos en el embalse El Limón ubicado en el norte de Argentina, conocido por las

proliferaciones recurrentes de cianobacterias. Cada muestreo se clasificó en cuatro niveles basados en la abun-

dancia de cianobacterias (células/ml): Nivel 1 (10 000-30 000); Nivel 2 (30 000-50 000); Nivel 3 (50 000-100 000)

y Nivel 4 (> 100 000). Para cada nivel, se construyeron redes de correlación de abundancias considerando todas

las especies.

Resultados: Se observó un patrón de disminución de correlaciones de abundancia estadísticamente significativas

a medida que aumentaba la intensidad de las proliferaciones: 219 correlaciones en el nivel 1; 144 en el nivel 2; 80

en el nivel 3 y solo 33 en el nivel 4. Las cianobacterias que forman proliferaciones mostraron tener poca corre-

lación con otras especies en todos los niveles, lo que podría estar asociado a cierta independencia con respecto

a la comunidad. Un aumento en la intensidad de la proliferación parece desconectar la red de correlaciones de

abundancia del fitoplancton.

Conclusión: El análisis de las redes de correlaciones de abundancias debería ser una herramienta valiosa para

comprender la dinámica y el desarrollo de las proliferaciones de cianobacterias, así como para identificar especies

clave en este proceso.

Palabras clave: embalse; fitoplancton; Argentina; ecología; manejo.

Revista de Biología Tropical, ISSN: 2215-2075, Vol. 72: e56487, enero-diciembre 2024 (Publicado Abr. 03, 2024)

can show artifacts (Feng et al., 2019; Lozano,

2022). Nevertheless, considering both abiotic

and biotic interactions is fundamental to a

better comprehension of CB and both need to

be assessed.

To test the possibility of using the corre-

lation of abundance networks as an early alert

tool for CB, we analyzed 4 levels of CB in a

tropical reservoir in Argentina.

MATERIALS AND METHODS

Study area: El Limón (22°6’12.29’’ S &

63°44’21.34’’ W) reservoir is in the Northern

province of Salta, Argentina. It covers approxi-

mately 100 ha with a capacity of 1.7 Hm3, and

an average depth of 4 m. The climate in the area

is distinctly tropical with high temperatures

during the dry season and an annual average

rainfall exceeding 970 mm (Arias & Bianchi,

1996). The average altitude of the region stands

at 550 m.a.s.l. in an area referred to as the “pied-

mont” or “transitional jungle”.

Sampling: Between June 2018 and January

2020, a total of 22 samples (in 1 l bottles) were

collected monthly in the El Limón reservoir at

the Secchi depth. Additionally, some bimonthly

samplings were conducted during the warmer

months, as they are associated with periods of

bloom development. The sampling site remai-

ned constant at the only access point to the

Reservoir, situated approximately 15 m from

the shoreline (Fig. 1).

Taxonomic phytoplankton counting: For

phytoplankton analysis, qualitative samples

were collected below the surface using a 30 µm

mesh net, that was dragged horizontally, and

fixed with 4% formaldehyde. Formaldehyde-

fixed samples were used solely for support

purposes in taxonomic identification. Quan-

titative analysis was carried out using samples

taken at the depth of a Secchi disk, fixed in

acidified Lugol’s solution, and stored at 4 °C

until analysis. After 24 h of sedimentation, cou-

nts were performed using combined chambers

Fig. 1. El Limón Reservoir and pictures of the study area. A. and B. Photographs of the El Limón Reservoir. C. Study area

map, the orange dot in the reservoir marks the sampling site.

4Revista de Biología Tropical, ISSN: 2215-2075 Vol. 72: e56487, enero-diciembre 2024 (Publicado Abr. 03, 2024)

on an inverted Zeiss L microscope, following

the method by Utermöhl (1958). Each sample

was counted to obtain less than a 20% error

for the most frequent species (Venrick, 1978).

The results were expressed in cells/ml. The

number of cells per filament was determined by

dividing the total filament length by the mean

cell length (N= 20). Organisms without cellular

content were excluded from the count. Species

were identified by capturing images using an

Axio Cam1Cc3 digital camera and utilizing

specialized references such as Komárek and

Anagnostidis (1999), Komárek and Anagnos-

tidis (2005), Komárek (2014), Komárková-Leg-

nerová (1969), Krammer and Lange-Bertalot

(1986), among others.

In this study, a CB was considered when

the abundance of at least one species of cya-

nobacteria exceeded 5 000 cells/ml. Based on

total cyanobacterial abundances, each sample

was classified into different categories or levels.

Four levels were pre-established based on total

abundances (cells/ml): Level 1 (10 000-30 000);

Level 2 (30 000-50 000); Level 3 (50 000-

100 000); and Level 4 (> 100 000).

Physical and Chemical Variables: In all

collections, a thermal profile of the reservoir

was conducted. In-situ measurements were

taken for temperature (°C), electrical conduc-

tivity (µs/cm), pH, and dissolved oxygen (D.O.)

(mg/l) using an Orion multiparameter sensor.

Additionally, turbidity was measured using a

HACH turbidimeter (NTU), and transparency

was assessed with a Secchi disk. Samples for

physical and chemical analyses were collected

using a Van Dorn sampler at the depth of 1-

Secchi disk and refrigerated until analysis. Total

and suspended solids (mg/l), true color, nitrates,

nitrites, and ammonium (mg N/l), soluble reac-

tive phosphorus (mg PRS/l), chemical oxygen

demand (mg O2 /l), alkalinity (mg CaCO3/l),

and hardness (mg CaCO3/l) were determined

in the laboratory following standardized APHA

techniques (APHA, 2005). Chlorophyll a con-

centration was measured using the modified

Scor-Unesco technique (Cabrera-Silva, 1984).

The trophic state of the reservoir was assessed

using the Carlson Trophic State Index (TSI)

based on chlorophyll a (Carlson, 1977).

Statistical Analysis: Using specific phyto-

planktonic abundances by level, Spearman

correlation coefficients were calculated using

InfoStat v.2008 (Di Rienzo et al., 2010) and

correlation of abundances networks were cons-

tructed using Cytoscape v.3.7.1. (Shannon et al.,

2003), considering only significant correlation

coefficients (P ≤ 0.01). The physical and che-

mical variables were compared between levels

using the Kruskal-Wallis non-parametric test

because variables did not meet the normality

and/or homogeneity requirements.

RESULTS

Based on the proposed classification of

levels, the 22 samplings were divided as follows:

4 categorized in level 1; 8 in level 2; 5 in level 3;

and 5 in level 4. Water temperature was statis-

tically different among the 4 levels (H = 9.22, P

= 0.0263). A trend of increasing temperatures

during the CB classified in the higher levels was

observed. Samplings classified in level 1 exhibi-

ted the lowest average air temperature (19.4 ±

6.01 °C), while those in level 4 recorded the hig-

hest average temperature in samplings (26.6 ±

2.84 °C). A similar pattern was observed in the

water temperature of the reservoir, with means

of 20.18 °C (± 4.95), 24.3 °C (± 2.84), 24.92 °C

(± 6.69), and 31.98 °C (± 4.15) in levels 1, 2, 3,

and 4, respectively.

Besides water temperature, the only water

chemistry parameter significantly different bet-

ween the levels was alkalinity (H = 8.84, P =

0.0314). Other parameters remained relatively

stable throughout the analyzed period (Table 1).

The Carlson trophic index was calculated

for all samplings based on chlorophyll a. The

overall state of the reservoir was found to be

mesotrophic, with only 3 eutrophic samplings

corresponding to samplings in levels 1 and 2 in

the wet season when the water level was 0.5 m

superior to the average (4 m). Regarding phyto-

planktonic species richness over the entire con-

sidered period, 162 species were identified. The

Revista de Biología Tropical, ISSN: 2215-2075, Vol. 72: e56487, enero-diciembre 2024 (Publicado Abr. 03, 2024)

most important group was the Chlorophyceae

with 65 spp., followed by Cyanobacteria with

51 spp. Additionally, 22 species of Bacillario-

phyceae, 11 of Euglenophyceae, 5 of Crypto-

phyceae, and 3 of Dinophyceae were identified.

Species richness was 66 spp. in level 1, 83

spp. in level 2, 56 spp. in level 3, and 42 spp. in

level 4. The highest richness of cyanobacteria

was observed in level 2 (41 spp.), followed by

level 1 (33 spp.), despite their lower abundances

compared to levels 3 and 4. Dominant cyano-

bacteria species in each level were Aphanocapsa

elachista in level 1 and Raphidiopsis medite-

rranea in levels 2, 3, and 4. The Venn diagram

in Fig. 2 illustrates the species that overlap

between different levels and those exclusive to

each one. 27 species were found at all levels.

Specifically, level 1 documented 9 exclusive spe-

cies, while 11, 3 and 1 were exclusive to levels 2,

3 and 4 respectively.

Total phytoplankton abundances showed

statistically significant differences (H = 19.1,

P = 0.0003). A sustained increase in total

phytoplankton abundance was observed at

Table 1

Average and standard deviations of physical and chemical variables for each level.

Varia ble Level 1 Level 2 Level 3 Level 4 Statistical differences

pH 7.53 ± 0.64 7.18 ± 0.41 7.45 ± 0.44 7.29 ± 0.72 (H = 1.63, P = 0.652)

E.C. (µS/cm) 610.4 ± 32.7 547.6 ± 122.8 457.4 ± 168.4 600.9 ± 62.1 (H = 3.23, P = 0.356)

Turbidity (NTU) 3.14 ± 1.67 7.26 ± 6.87 3.88 ± 2.13 4 ± 1.78 (H = 4.51, P = 0.211)

Alkalinity (mg CaCO3/l) 132.3 ± 92.4 86.7 ± 14.4 153.7 ± 68.7 205.1 ± 27.1 (H = 8.84, P = 0.031)

Hardness (mg CaCO3/l) 217.9 ± 50.2 434.0 ± 549.4 275.2 ± 114.1 164.6 ± 51.4 (H = 6.01, P = 0.111)

N/P 17.6 ± 6.6 24.0 ± 30.3 16.42 ± 4.0 9.7 ± 7.4 (H = 2.92, P = 0.404)

D.O. (mg O2/l) 9.35 ± 1.40 8.77 ± 2.39 8.98 ± 1.52 8.04 ± 1.82 (H = 1.10, P = 0.777)

T (°C) 20.1 ± 4.9 24. ± 2.84 24.92±6.69 31.98 ± 4.15 (H = 9.22, P = 0.026)

Fig. 2. Phytoplanktonic richness in the different levels. Venn diagram shows the species shared between levels. Pie charts

show the species classification in broad taxonomic groups.

6Revista de Biología Tropical, ISSN: 2215-2075 Vol. 72: e56487, enero-diciembre 2024 (Publicado Abr. 03, 2024)

higher levels. In this regard, the phytoplankton

abundance in level 4 was 1 117 % higher than

the average total abundance recorded in level

1, corresponding level 4 to the most inten-

se blooms. Cyanobacterial mean abundances

values were: 18 693 (± 5 787 cells/ml), 38 188

(± 6 968 cells/ml), 62 924 (± 15 913 cells/ml),

and 261 260 (± 262 487 cells/ml) for levels 1, 2,

3, and 4, respectively.

Abundance correlations for all phytoplank-

ton species present in each level were analyzed

considering only highly significant correlations

(P ≤ 0.01). A pattern of decreasing significantly

correlated records at higher levels was obser-

ved: 219 significant correlations were recorded

in level 1; 144 in level 2; 80 in level 3, and

only 33 in level 4 (Fig. 3). Correlations abun-

dances between species were very different

between bloom levels; being mostly positive

correlations. With the dominance of cyano-

bacteria, correlation abundances decreased

drastically, both between non-cyanobacteria

Fig. 3. Correlation abundances networks by level. Black edges represent positive correlations while red ones denote

negative correlations. Circle sizes are proportional to specific abundances. Colors show groups: blue: Cyanobacteria, green:

Chlorophyceae, yellow: Bacillariophyceae, turquoise: Euglenophyceae, dark brown: Dinophyceae, light brown: Ochrophyta,

red: Xantophyta. * The dominant species (higher abundances) are highlighted in the figure. The black lines in the charts

represent positive correlations, and the red lines represent negative correlations.

Revista de Biología Tropical, ISSN: 2215-2075, Vol. 72: e56487, enero-diciembre 2024 (Publicado Abr. 03, 2024)

phytoplankton species, and between cyanobac-

teria ones. In general, dominant cyanobacteria

species at each level were always poorly corre-

lated with all other species.

DISCUSSION

Correlation abundance networks aid

in understanding the relationships within

the phytoplanktonic community and could

be a powerful tool for monitoring cyano-

bacteria blooms.

The trophic state of the El Limón reservoir

was predominantly mesotrophic throughout

the period, which aligns with the overall state

of reservoirs in Argentina (O’Farrell et al., 2019;

Amé et al., 2003; Bazán et al., 2005). Nutrient

concentrations, mainly dissolved nitrogen and

soluble reactive phosphorus, showed no sig-

nificant differences over the analyzed cycle or

between proposed classification levels. Based

on these results, we can infer that the differen-

ces in the intensity of the blooms may be mostly

promoted by temperature, which did show

differences between levels, and helped by the

biological interactions of the bloom-forming

species with the rest of the community.

Cyanobacterial blooms have been recu-

rrent in the El Limón reservoir during the study

period, highlighting the issue and rendering

the reservoir a risk-prone environment due to

its use for water consumption. Over time, these

blooms became increasingly intense. Previous

studies conducted in tropical reservoirs exhibit

similarities to the results of the present work,

with a predominant representation of Aph. gra-

cile, R. raciborskii, M. flos-aquae, Pseudanabae-

na spp., R. mediterranea, and Dolichospermum

spp. (Harke et al., 2016). These genera have

been frequently observed in Argentina as well

as in reservoirs in the central and Southern

regions of Brazil (Echenique et al., 2006; Salus-

so & Moraña; 2018), despite existing mor-

phohydrological differences between them

(Moschini et al., 2009; Sant’Anna et al., 2007).

These similarities in cyanobacterial commu-

nity components based on climatic conditions

might indicate a geographic spread of blooms

at a regional level (Bittencourt-Oliveira et al.,

2014), especially of invasive species that have

expanded their current dispersal range, such as

Cylindrospermopsis (Cires & Ballot, 2016).

The analysis of correlation abundance net-

works at different bloom intensities allows us

to delve into the effect of blooms on the phyto-

plankton community and to formulate biolo-

gical hypotheses of bloom-forming algae and

cyanobacteria. Despite decreasing species rich-

ness during intense blooms, the relationships

among phytoplankton species also declined.

But what are the possible ecological impli-

cations of correlated abundances? positively

correlated abundances could be associated with

co-aggregation, cross-feeding, co-colonization,

niche overlap, cooperation, or facilitation, while

negative relationships could be linked to com-

petition or amensalism (Deng et al., 2012; Faust

& Raes, 2012). In the reservoir, we predomi-

nantly observed positive correlations, which

might indicate the lack of net competition for

resources. The prevalence of positive corre-

lations could imply that all species respond

similarly to external stimuli, such as environ-

mental factors, thus their population growth

is constrained by the same parameters. Given

the predominantly mesotrophic to eutrophic

state of the reservoir, it can be assumed that

macronutrient requirements (N and P) were

more than fulfilled during all sampling dates.

The overwhelming positive correlations across

all levels could be also explained by the possible

cooperation leading to coupling and positive

feedback among phytoplankton species, enhan-

cing overall metabolic efficiency within the

community (Coyte et al., 2015).

Remarkably, predominant cyanobacterial

species exhibited limited interconnectivity in

the correlated abundance networks observed at

low bloom levels; appearing to have maintained

a degree of isolation from the broader commu-

nity in the most severe bloom, they were all of

them practically completely disconnected. For

example, Raphidiopsis mediterranea in level 3

was correlated only with 2 species, Crucigenia

tetrapedia and Cyclotella sp., while in level 4, it

was correlated only with Scenedesmus spinosus.

8Revista de Biología Tropical, ISSN: 2215-2075 Vol. 72: e56487, enero-diciembre 2024 (Publicado Abr. 03, 2024)

Given the low number of correlations in these

levels, these correlations might well be ran-

dom and lack biological significance. The fact

that the primary bloom-forming species beco-

mes completely disconnected from the phyto-

plankton network could indicate its capacity to

dominate the phytoplankton network (Lozano,

2022). Conversely, strongly correlated species in

a network respond similarly to environmental

conditions, probably limiting the capability of

each one to dominate. The disconnection from

the community could increase the probability

of successful bloom, as it remains “indepen-

dent” of the network. Abundance correlation

networks could be useful in identifying bloom-

forming species, as their ecological behavior is

key to their dominance. This study is explora-

tory, seeking new applications for the tool, so

it is advisable to continue refining it and use

larger sample sizes.

Ethical statement: the authors declare

that they all agree with this publication and

made significant contributions; that there is

no conflict of interest of any kind; and that we

followed all pertinent ethical and legal proce-

dures and requirements. All financial sources

are fully and clearly stated in the acknowled-

gments section. A signed document has been

filed in the journal archives.

ACKNOWLEDGMENTS

The authors would like to thank the

reviewers and editors who participated in the

review process of this article for their time and

dedication. To CONICET and the National

University of Salta (UNSa) for the financing of

this study.

REFERENCES

Amé, M., Diaz, M., & Wunderlin, D. (2003). Occurrence of

toxic cyanobacterial blooms in San Roque reservoir

(Córdoba, Argentina): a field and chemometric study.

Environmental Toxicology, 18(1), 192–201.

APHA. (2005). Standard methods for the examination of

water and wastewater. APHA.

Arias, M., & Bianchi, A. (1996). Estadísticas climatológicas

de la Provincia de Salta. Dirección de Medio Ambien-

te y Recursos Naturales, Argentina.

Bazán, R., Corral, M., Pagot, M., Rodríguez, A., Oroná, C.,

Rodríguez, M. I., & Busso, F. (2005). Teledetección

y modelado numérico para el análisis de la calidad

de agua del embalse Los Molinos, Córdoba, Argen-

tina. Revista Ingeniería Hidráulica en México, 20(2),

121–135.

Berry, D., & Widder, S. (2014). Deciphering microbial

interactions and detecting keystone species with co-

occurrence networks. Frontiers in Microbiology, 5(2),

219–225.

Bittencourt-Oliveira, M., Carmo, D., Piccin-Santos, V.,

Moura, A., Aragão-Tavares, N., & Cordeiro-Araújo,

M. (2014). Cyanobacteria, microcystins and cylin-

drospermopsin in public drinking supply reservoirs

of Brazil. Anais da Academia Brasileira de Ciências,

86(1), 297–310.

Cabrera-Silva, S. (1984). Estimación de clorofila a y feopig-

mentos. Una revisión metodológica. UNESCO.

Carlson, R. (1977). A trophic state index for lakes. Limnolo-

gy and Oceanography, 22(1), 361–369.

Cheung, M., Liang, Y., & Lee, J. (2013). Toxin-producing

cyanobacteria in freshwater: a review of the pro-

blems, impact on drinking water safety, and efforts

for protecting public health. Journal of Microbiology,

51(2), 1–10.

Cires, S., & Ballot, A. (2016). A review of the phylogenetic,

ecology and toxin production of bloom-forming

Aphanizomenon spp. and related species within the

Nostocales (Cyanobacteria). Harmful Algae, 54(1),

21–43.

Coyte, K. Z., Schluter, J., & Foster, K. R. (2015). The ecology

of the microbiome: networks, competition, and stabi-

lity. Science, 350(6261), 663–666.

D’Amen, M., Mod, H. K., Gotelli, N. J., & Guisan, A. (2018).

Disentangling biotic interactions, environmental fil-

ters, and dispersal limitation as drivers of species co-

occurrence. Ecography, 41(8), 1233–1244.

Deng, Y., Jiang, Y. H., Yang, Y., He, Z., Luo, F., & Zhou, J.

(2012). Molecular ecological network analyses. BMC

Bioinformatics, 13(1), 1–20.

Di Rienzo, J., Balzarini, M., Gonzalez, L., Casanoves, F.,

Tablada, M., & Robledo, C. W. (2010). InfoStat: soft-

ware para análisis estadístico. InfoStat.

Echenique, R., Gianuzzi, L., & Ferrari, L. (2006). Drinking

water: problems related to water supply in Bahía

Blanca, Argentina. Acta Toxicológica Argentina, 14(2),

23–30.

Revista de Biología Tropical, ISSN: 2215-2075, Vol. 72: e56487, enero-diciembre 2024 (Publicado Abr. 03, 2024)

Faust, K., & Raes, J. (2012). Microbial interactions: from

networks to models. Nature Reviews Microbiology,

10(8), 538–550.

Feng, K., Zhang, Y., He, Z., Ning, D., & Deng, Y. (2019).

Interdomain ecological networks between plants

and microbes. Molecular Ecology Resources, 19(6),

1565–1577.

Harke, M. J., Steffen, M. M., Gobler, C. J., Otten, T. G.,

Wilhelm, S. W., Wood, S. A., & Paerl, H. W. (2016). A

review of the global ecology, genomics, and biogeo-

graphy of the toxic cyanobacterium, Microcystis spp.

Harmful Algae, 54(2), 4–20.

Kokociński, M., Dziga, D., Antosiak, A., & Soininen, J.

(2021). Are bacterio-and phytoplankton community

compositions related in lakes differing in their cyano-

bacteria contribution and physico-chemical proper-

ties? Genes, 12(6), 850–855.

Komárek, J. (2014). Modern classification of cyanobacteria.

In N. V. Sharma, A. K. Rai, & L. J. Stal (Eds.), Cyano-

bacteria: An economic perspective (pp. 21–39). John

Wiley & Sons, Ltd.

Komárek, J., & Anagnostidis, K. (1999). Cyanoprokaryota,

1. Teil: Chroococcales. In H. Ettl., G. Gärtner, H.

Heynig, & D. Mollenhauer (Eds.), Subwasserflora von

Mitteleuropa. Spektrum, Akademischer Verlag.

Komárek, J., & Anagnostidis, K. (2005). Cyanoprokaryota.

2. Oscillatoriales. In B. Büdel, G. Gärtner, L. Krienitz,

& M. Schagerl (Eds.), Süβwasserflora von Mitteleuro-

pa. Spektrum, Akademischer Verlag.

Komárková-Legnerová, J. (1969). The systematics and

ontogenesis of the genera Ankistrodesmus Corda and

Monoraphidium gen. nov. In B. Fott (Ed.), Studies in

Phycology (pp. 75–144). Schweizerbart’sche.

Krammer, K., & Lange-Bertalot, H. (1986). Bacillarío-

phyceae, 1. Teil: Naviculaceae. In H. Ettl, J. Gerloff, H.

Heynig, & D. Mollenhauer (Eds.), Süβwasserflora von

Mitteleuropa. Gustav Fischer Verlag.

Kudela, R., Berdalet, E., Enevoldsen, H., Pitcher, G., Raine,

R., & Urban, E. (2017). “GEOHAB–The global eco-

logy and oceanography of harmful algal blooms pro-

gram: motivation, goals, and legacy”. Oceanography,

30(1), 12–21.

Lozano, V. L. (2022). Hidden impacts of environmen-

tal stressors on freshwater communities could be

revealed at lower concentrations by correlation of

abundances network analyses: An example with her-

bicides glyphosate, 2, 4-D, and their mixtures. Austral

Ecology, 47(5), 1144–1153.

Lu, L., Tang, Q., Li, H., & Li, Z., (2022). Damming river

shapes distinct patterns and processes of planktonic

bacterial and microeukaryotic communities. Environ-

mental Microbiology, 24(4), 1760–1774.

Moschini, C., Bortoli, E., Pinto, P., Nishimura, L., Freitas,

M., Pompeo, F., & Dorr, F. (2009). Cyanobacteria and

cyanotoxin in the Billings reservoir (São Paulo, SP,

Brazil). Limnetica, 28(1), 273–282.

O’Farrell, I., Motta, C., Forastier, M., Polla, W., Otaño, S.,

Meichtry, N., Devercelli, M., & Lombardo, R. (2019).

Ecological meta-analysis of bloom-forming plank-

tonic Cyanobacteria in Argentina. Harmful Algae,

83(1), 1–13.

O’Neil, J. M., Davis, T. W., Burford, M. A., & Gobler, C. J.

(2012). The rise of harmful cyanobacteria blooms:

the potential roles 584 of eutrophication and climate

change. Harmful Algae, 14, 313–334.

Paerl, H. W., & Otten, T. G. (2013). Harmful cyanobacterial

blooms: causes, consequences, and controls. Microbial

Ecology, 65, 995–1010.

Paerl, H. W., & Otten, T. G. (2016). Duelling “CyanoHABs”:

unravelling the environmental drivers controlling

dominance and 576 succetavssion among diazotro-

phic and non-N2-fixing harmful cyanobacteria. Envi-

ronmental Microbiology, 18, 316–324.

Rüger, L., Feng, K., Dumack, K., Freudenthal, J., Chen, Y.,

Sun, R., & Bonkowski, M. (2021). Assembly patterns

of the rhizosphere microbiome along the longitudinal

root axis of maize (Zea mays L.). Frontiers in Micro-

biology, 12, 614501.

Salusso, M. M., & Moraña, L. (2018). Comparative reser-

voir limnology in Juramento (Salta) and Salí-Dulce

(Tucumán) Basins in Argentina. Revista de Biología

Tropical, 66(1), 415–427.

Sant’Anna, C., Silva, M., Carvhalom, M., Gelmego, M., &

Azevedo, M. (2007). Planktic cyanobacteria from

upper Tiete basin reservoir, SP, Brasil. Brazilian Jour-

nal of Botany, 30(1), 1–17.

Shannon, P., Markiel, A., Ozier, O., Baliga, N. S., Wang, J.

T., Ramage, D., Amin, N., Schwikowski, B., & Ideker,

T. (2003). Cytoscape: a software environment for inte-

grated models of biomolecular interaction networks.

Genome Research, 13(11), 2498–2504.

Utermöhl, H. (1958). On the perfecting of quantitative

phytoplankton method. International Association of

Theory and Applications of Limnology Proceedings,

9, 1–38.

Venrick, E. L. (1978). How many cells to count? Phytoplank-

ton Manual, 6, 167–180.

World Health Organization. (2003). Guidelines for Safe

Recreational Water Environments, Volume 1: Coastal

and Fresh Waters. World Health Organization.